The Surface Water and Ocean Topography (SWOT) mission provides spatially continuous observations of river water surface elevation (WSE), offering new opportunities to complement traditional gauge-based monitoring. This study investigates the integration of SWOT- derived WSE with in-situ observations from the United States Geological Survey (USGS) for improved river characterization over the Willamette River. A comprehensive quality-control framework is applied to SWOT data to remove ice-affected observations, low-quality retrievals, geolocation errors, swath-edge artifacts, and statistical outliers, ensuring robust open-water WSE estimates. The filtered dataset is analyzed along different reaches and compared with stage and streamflow observations from USGS gauges at Corvallis, Albany, Salem, and Jefferson. Initial diagnostics reveal strong seasonal behavior, with peak flows occurring during winter through late spring, and indicate both linear and nonlinear relationships between stage and discharge. To model these relationships, a Bayesian hierarchical framework is developed in which SWOT-derived WSE at each node is expressed as a function of upstream and downstream USGS gauge stage and discharge. The model follows a three-layer structure: (i) a data layer where WSE is modeled using a parametric distribution with time-varying parameters; (ii) a process layer linking these parameters to hydrologic covariates based on river reach connectivity; and (iii) a prior–likelihood layer enabling uncertainty quantification and information sharing across nodes. Posterior distributions are estimated using Markov Chain Monte Carlo (MCMC), allowing probabilistic inference of river stage dynamics. Model performance is evaluated at both node and reach scales. Results demonstrate strong agreement between modeled and SWOT-observed WSE, with Pearson correlation (R) values consistently high (generally > 0.9 across most nodes) and Nash–Sutcliffe Efficiency (NSE) indicating good predictive skill for a majority of nodes. Some localized reductions in NSE are observed at specific nodes, likely associated with increased hydraulic complexity or measurement uncertainty, but overall performance remains robust across the reach. These findings highlight the potential of combining SWOT altimetry with ground-based observations within a Bayesian framework to enhance river monitoring, capture spatial variability, and provide probabilistic estimates of river state. The proposed approach offers a scalable pathway for integrating next-generation satellite observations into hydrologic modeling and flood assessment systems.